What is our primary use case?
We're using OpenLegacy to monetize our mainframe assets. I've been doing this for 40-plus years. I was working in diagnostics since CICS and Mainframer. Then I moved to distribution back in the nineties. Everybody has tried to get away from those mainframe applications and assets as much as possible, but the reality is you literally can't. Our bank was bought by this $500 billion bank working with Hogan. I was a Hogan expert back in the eighties and nineties. It's an early-1990s mainframe. Once you're tied up with that kind of stuff, you're not going to easily get off it. I had a boss who always used to say we should be able to get off a mainframe application in two to three weeks. He got fired.
There are many people who will try and get you off a mainframe and say that we can port the data. Of course, you can import the data, but here's what people fail to understand about a mainframe. IMS, CICS-transaction, or shared databases — whether it's by IMS Db2 or even VSAM— have been around for 40 years. So that means that, for all intents and purposes, the mainframe is a big PC with a lot of DLLs. And those DLLs allow you to jump from one application seamlessly to another application.
So over a 40-year period, it's likely that nobody ever built a component-based architecture on a mainframe, or they may have started with a component-based architecture on a mainframe, but you have data sharing through shared data sets where you have program-to-program jumps from one application to another seamlessly. When someone says they can pull an application out, they failed to take into account 15 other applications that have dependencies.
That's the challenge with a project like this. It's like that Jenga game where you pull stuff out. You need technology that allows you to have a bidirectional call in and call out. You have to look for tooling that is flexible within the constraints that you have for that mainframe in the portfolio. Our project went south, and it was a half a billion-dollar project.
It might be considered to be a hybrid deployment. On the Microsoft service side, we're running an AWS application in our own tenant. But it's still an AWS public cloud. Obviously, there's the kernel and pieces of code that run on the mainframe. And those pieces that are on the mainframe would be on-prem. So I would call it a hybrid deployment.
What is most valuable?
The biggest advantage of OpenLegacy was how simple the technology was. We were able to build out the OpenLegacy parts very quickly. We put together a couple hundred APIs in six months.
What needs improvement?
I'd like to see OpenLegacy develop its low-code/no-code (LCNC) solutions. They've expanded somewhat their horizons for integration beyond mainframe CICS, which is their sweet spot. They have some tooling in that area, but it's not as good as it needs to be.
OpenLegacy handles the bread-and-butter TP monitoring stuff, but I am working for one of the six banks in the United States still using the Hogan mainframe, which has a slightly proprietary mechanism. But OpenLegacy currently doesn't have a connector for Hogan. So it would help if they could build one. That would appeal to financial institutions that still use Hogan, like US Bank, Wells Fargo, KeyBank, and Vanguard.
For how long have I used the solution?
We've been using OpenLegacy for two years.
How was the initial setup?
We were able to execute mainframe enablement within six months. We had all that built.
What about the implementation team?
We partnered with FIS Global, one of the largest and most prominent banking software vendors. They had solutions that actually work in production. Our leadership at the time would not look at those solutions that were working. Instead, they wanted to partner with a vendor to build a new microservice-based platform because we were imposing that on our vendor partner. Of course, they'll do anything for money, but they couldn't deliver on schedule.
That's really where we're the problem began and ended. We were building our stuff and were on time with our delivery, but our vendor partner was constantly missing their deadline, and you're burning $20 million dollars a month. It doesn't take very many months before you've wasted a lot of money without any progress.
What's my experience with pricing, setup cost, and licensing?
OpenLegacy charges you based on the API. I'm not exactly sure, but I think the final negotiated price was something north of $100,000, but less than $200,000 for upwards of a thousand APIs. So it was a very reasonable price point. IBM pricing is built around the Altair size or the size of the machine. If you're running a big machine efficiently, you pay a penalty for running a big machine. So the IBM pricing tends to be not competitive.
Which other solutions did I evaluate?
We chose OpenLegacy over the IBM product for a few reasons. When it comes to monetizing existing assets — mainframe assets in particular — there are two use cases. One has been solved for years, and that's the ability to expose a mainframe asset as an API. Several technologies can do this, such as the old MQ CICS bridge and IMS Connect.
But the real challenge has been the ability of mainframe assets or IE programs to consume a microservice API. That's why we gravitated toward OpenLegacy. I know IBM can also do the same thing, but here's the second. The problem is most of the dev teams don't want to touch the mainframe. The mainframe is like a mysterious box that your grandfather worked on.
IBM tooling requires a level of knowledge that most contractors don't have. OpenLegacy involves a very minimal effort on the part of the mainframe team. Connecting is very simple because it connects straight into the CICs applications. It was simply a matter of creating a sub-program and passing the right COBALT copybook out to something called the API caller. There is a bit of work on the mainframe, but it's very contained. At the same time, the other solutions required too much mainframe knowledge for a project where you have a lot of offshoring. We wanted to use primarily non-specialized IT people working with Java or .NET because of those microservices. That was the primary driver: much less complexity. The cost was a factor, but it mainly was the complexity.
OpenLegacy caters to a specific niche, which is one of the reasons why we went with them. We were also looking at MuleSoft, which is a fabulous company that makes a lot of things easy. But even MuleSoft couldn't jump to the mainframe like OpenLegacy. That was really an eye-opener. While MuleSoft is a solid solution, you still have a niche problem with connectors, which OpenLegacy was able to solve. But ultimately we went for an open-source solution and then addressed the niche requirements with solutions like OpenLegacy.
What other advice do I have?
I would give OpenLegacy an eight out of 10. Almost no solution gets a 10. The reason for the eight is that it actually solved the problems that we had for 200-plus APIs very quickly. It was very impressive. There were some glitches, but OpenLegacy was extremely responsive. They don't get a 10 because we still couldn't employ their low code/no code.
Which deployment model are you using for this solution?
Hybrid Cloud
Disclosure: I am a real user, and this review is based on my own experience and opinions.
